No abstract

Pruning methods for feed-forward artificial neural networks trained by the cascade-correlation learning algorithm are proposed. The cascade-correlation algorithm starts with a small network and dynamically adds new nodes until the analyzed problem has been solved. This feature of the algorithm removes the requirement to predefine the architecture of the neural network prior to network training. The developed pruning methods are used to estimate the importance of large sets of initial variables for quantitative structureactivity relationship studies and simulated data sets. The calculated results are compared with the performance of fixed-size back-propagation neural networks and multiple regression analysis and are carefully validated using different training/test set protocols, such as leave-one-out and full cross-validation procedures. The results suggest that the pruning methods can be successfully used to optimize the set of variables for the cascade-correlation learning algorithm neural networks. The use of variables selected by the elaborated methods provides an improvement of neural network prediction ability compared to that calculated using the unpruned sets of variables.

We investigated the applications of back propagation artificial neural networks (ANN) for a small dataset analysis in the field of structure-activity relationships. The derivatives of carboquinone were used as an example. It's been found that in this case the use of the same neural network results in unambiguous classification of new molecules. Predictions can be improved with statistical analysis of independent prognosis sets. We suggest that the sign criterion be used as a classification rule. We also compared neural networks with FALS and ALS in leave-one-out prediction. ANN applied to the same dataset has shown the same predictive ability as ALS but poorer than FALS.

No abstract